Infective endocarditis: classification and clinical forms

Article written and reviewed by Dr Luca Moretti

Infective endocarditis includes infections that share an endocardial location but differ in substrate, anatomy, organism and acquisition setting. Classification translates these differences into decisions: which microorganisms to investigate and cover initially, which imaging to use, when to remove implanted material and which complications to monitor. A complete definition may therefore be, for example, healthcare-associated native mitral-valve endocarditis caused by Staphylococcus aureus, or late prosthetic aortic-valve endocarditis with negative blood cultures. Each element adds information not contained in the others.

The general monograph develops the shared diagnostic and treatment pathway; dedicated pages examine native-valve disease, prosthetic-valve disease, right-sided disease, cardiac devices, transcatheter valves, blood culture-negative disease and fungal forms. These categories overlap: a patient may simultaneously have a prosthesis, a device and a fungal pathogen, or both right- and left-sided infection. Classification should represent this complexity without forcing every case into a single category.

The proportions of the different forms depend on the population. In centers treating many older adults, prosthetic-valve recipients and dialysis patients, healthcare-associated and device-related infections have greater weight; in populations with injection drug use, right-sided locations are more common, without excluding left-sided disease. Contemporary registries describe a disease that is often complex and burdened by comorbidity. However, the frequency observed in a referral center is also affected by transfers for complications and surgery: it cannot be interpreted as the universal distribution of disease or as an individual probability without considering context.

Etiology, pathogenesis and pathophysiology

On a native valve, a surface altered by degeneration, malformation or haemodynamic injury favors adhesion to platelet-fibrin deposits. Some pathogens, particularly S. aureus, can infect a valve without a known pre-existing lesion. The absence of implanted material removes one possible reservoir but does not eliminate abscesses, destruction or persistent bacteremia. The clinically useful distinction is among a limited vegetation, leaflet damage and invasion beyond the annulus: three conditions that may require very different strategies despite all meeting the definition of native-valve endocarditis.

A prosthetic valve introduces artificial surfaces, sutures and interfaces in which biofilm can form. Infection may involve the leaflets, annulus and surrounding tissues, with loss of support, dehiscence or abscess. Time from implantation helps assess perioperative contamination and subsequent colonization, but the early-versus-late category does not replace healthcare history and microbiology. Temporal thresholds used in studies are not all identical. An old prosthesis can be infected by a resistant healthcare-associated pathogen, whereas a relatively early infection does not automatically prove transmission in the operating room.

A transcatheter valve has a frame and relationships with native tissues that differ from those of a surgical prosthesis. Vegetations may localize on the leaflets, stent or adjacent structures; artifacts can sometimes make the lesion less conspicuous. Advanced age and comorbidities, which often motivated the initial choice of TAVI, influence tolerance of infection and feasibility of reintervention. The previous decision in favor of a transcatheter procedure does not, however, by itself constitute a permanent declaration of inoperability: the new anatomical problem must be reassessed.

In infections involving cardiac implantable electronic devices, the pocket, generator, transvenous leads and valves are connected but distinguishable sites. A local infection can extend along the system, whereas bacteremia can seed it hematogenously. A mass on a lead is not necessarily infected, and a negative transesophageal study does not exclude device infection. Persistence of colonized material explains why control often requires complete extraction rather than generator replacement alone or prolonged antibiotic therapy. The need for pacing and the need to eradicate the focus should be planned together.

The distinction between the left and right heart mainly anticipates consequences. Mitral and aortic lesions can cause left-sided heart failure, perianular extension and systemic emboli; tricuspid lesions cause regurgitation and pulmonary dissemination. Right-sided infection can compromise respiration and circulation even without severe left ventricular dysfunction. Injection drug use, catheters and transvenous leads are relevant but not exclusive contexts; congenital heart disease and conduits may further modify sites and flows. The other valves must still be assessed because distribution is not always unilateral.

The acquisition setting distinguishes community-acquired from healthcare-associated disease, including healthcare exposures outside hospital. Dialysis, infusions, chronic vascular access and recent admissions may point toward resistant staphylococci or other organisms selected by the healthcare environment. Microbiological classification instead adds virulence and susceptibility: S. aureus, streptococci, enterococci, Gram-negative organisms and fungi have different therapeutic requirements. Empirical treatment arises from the intersection of these elements; it should not be inferred solely from the place where the first fever was measured.

Blood culture-negative endocarditis is not a uniform pathogenetic category. After antibiotics, the organism may be a common staphylococcus or streptococcus rendered non-cultivable; without antimicrobial exposure, specific testing and organisms that are difficult to recover become more important. Fungi and intracellular pathogens require different pathways, while a sterile mass remains an alternative diagnosis. This distinction determines both investigations and empirical treatment: managing all negative cultures with the same regimen confuses a limitation of the method with a biological identity.

A further distinction concerns active infection and sequelae. After cure, an organized vegetation or a damaged valve may persist without bacteremia or ongoing invasion. Conversely, an abscess may remain active even after blood cultures become negative. Classification should therefore include microbiological response, anatomical evolution and haemodynamic tolerance. This dynamic interpretation prevents prolonging drugs for a mass that is already sterile or regarding reduction in its size as sufficient while a mechanical defect continues to progress.

Clinical manifestations

Fever, fatigue, weight loss and signs of inflammation can occur in all forms, but their speed and intensity reflect the organism and host more than the category name itself. A native valve can be destroyed within days, whereas a prosthesis may have a subacute infection with modest symptoms. In older adults, delirium, anorexia and loss of independence may precede a recognizable infectious syndrome. The threshold for investigation should reflect risk and exposures, not a requirement to observe high fever or all classic peripheral signs.

In the left heart, dyspnea and heart failure may indicate acute regurgitation. Echocardiography showing non-dilated chambers is not reassuring when the lesion is recent and pressures are elevated. In a prosthetic-valve recipient, a new leak, altered valve sounds or haemodynamic deterioration require comparison with previous studies. In TAVI recipients, dyspnea and functional decline may mistakenly be attributed to frailty or pre-existing heart failure. A change in the clinical trajectory is often more useful than any isolated symptom.

In the right heart, respiratory symptoms may predominate: pleuritic pain, cough, hemoptysis, peripheral infiltrates and fever can result from septic emboli. A sequence of presumed pneumonias in a patient with bacteremia or venous access requires investigation of a cardiac source. Ascites and systemic congestion indicate the consequences of regurgitation or right-sided damage but may appear later. A systemic embolic event in a picture believed to be exclusively right-sided requires investigation for left-sided involvement, a shunt or another source.

In a patient with a device, the examination looks for pain, swelling, discharge and pocket erosion, without limiting inspection to the wound. The absence of local signs is compatible with hematogenous lead infection. The presence of a device also changes the significance of persistent or recurrent bacteremia, especially from organisms capable of adhering to biomaterials. Assessment should also document pacing dependence and the type and age of the leads because these data influence extraction and reimplantation planning.

Culture-negative forms may present with emboli, valvular insufficiency or immunological phenomena, and fungal forms with large vegetations and dissemination; however, no clinical pattern replaces organism identification. A history of antibiotics, animal contact, occupational exposures, housing conditions and immunosuppression guides targeted testing. Vertebral, splenic or joint pain suggests a secondary focus that must be incorporated into the plan. Useful classification therefore also includes complications already present because these often determine urgency more than the still-unknown organism.

Investigations and diagnosis

The common foundations are blood cultures and echocardiography. Sampling before antibiotics and adequate blood volume improve yield; in unstable patients, collection should be rapid. Transthoracic echocardiography describes the lesion and cardiac function, but how reassuring a negative result is depends on image quality and substrate. A prosthesis or device may obscure the focus, whereas a large tricuspid vegetation may be clearly visible. The choice of transesophageal echocardiography and complementary examinations should begin with the specific limitation of the initial study.

In native left-sided disease, transesophageal echocardiography further evaluates vegetations, perforations and perianular complications, even after a positive TTE. In clearly defined native right-sided disease with high-quality images, it may not always be necessary if there are no additional questions. When suspicion persists despite a negative or uncertain study, repeating imaging generally within 5-7 days, or earlier if the condition worsens, looks for evolution and initially invisible lesions. A negative PET study has limited ability to exclude small native-valve vegetations and should not replace this reassessment.

In prosthetic-valve disease, echocardiography, CT and metabolic imaging are often complementary. CT defines perianular extension and anatomical relationships; PET can support infection when anatomy is difficult to interpret on echocardiography. Uptake pattern, postoperative interval and implanted materials should be considered to distinguish infection from sterile inflammation. New paravalvular regurgitation or dehiscence increases suspicion, but comparison with the postoperative baseline is necessary: a previously known leak does not demonstrate a new infection.

In device infections, microbiology and pocket signs can be decisive even when no vegetation is seen on the leads. Conversely, an isolated mass does not prove infection. Metabolic imaging and a search for other sources help in uncertain cases; extracted material should be analyzed with attention to contamination and to passage through an infected pocket. Assessment does not end with identification of a pathogen: it must establish which components are involved and whether concomitant infected valvular disease requires surgery.

The 2023 Duke-ISCVID criteria and 2023 ESC criteria update classification but retain differences. Duke-ISCVID includes additional typical organisms in the presence of prosthetic material, molecular diagnostics and specific surgical evidence; for prosthetic PET it also considers the interval from implantation. ESC accepts pathological prosthetic uptake patterns as major criteria regardless of that time limit, with expert interpretation. The most favorable elements of the two systems should not be selected to build a hybrid diagnosis. The report should state which system is being applied and which evidence supports each category.

Clinical combinations in the 2023 Duke-ISCVID classification:

• Definite: two major criteria, or one major and three minor criteria, or five minor criteria.
• Possible: one major and one minor criterion, or three minor criteria, in the absence of requirements for definite or rejected endocarditis.
• Pathological confirmation is a separate pathway; multiple concordant tests from the same imaging category are not independent major criteria.

Targeted microbiology completes the work-up when blood remains culture-negative or the organism has particular characteristics. Surgical tissue should be divided for culture, histology and molecular testing before fixation. Serology and PCR should be selected according to exposures and probability, avoiding interpretation of an isolated positive result as universal proof of active disease. Extracardiac imaging looks for lesions that explain symptoms or change treatment and surgery: a splenic abscess, spondylodiscitis and cerebral hemorrhage pose different problems. Failure to meet the definite category does not demonstrate sterile endocarditis.

Treatment and prognosis

Antimicrobial selection depends on the intersection between organism and substrate. Antibiotics active against a susceptible streptococcus may be appropriate for both native and prosthetic valves but with different durations; staphylococcal prosthetic-valve disease adds specific problems of biofilm and combination therapy. Healthcare-associated forms require empirical coverage consistent with resistance patterns and exposures. As soon as a reliable microbiological diagnosis is available, therapy is narrowed and dosed to achieve adequate exposure. Classification should justify these choices, not maintain broad empirical treatment after the question has been resolved.

Abbreviated regimens are not transferable across categories. Some streptococcal native-valve or right-sided forms may qualify for short regimens only under strict criteria; a prosthesis, abscess or metastatic focus changes applicability. Oral continuation therapy after an intravenous phase is instead a matter of clinical and pharmacological selection, not a synonym for short-course treatment. The POET trial supports defined combinations in stable patients with selected left-sided endocarditis, including some prosthetic valves; it does not support extension to uncontrolled infection, uncertain absorption or unrepresented microorganisms.

Valve surgery addresses heart failure, uncontrolled infection and embolic risk. In prosthetic-valve disease, dehiscence and perianular invasion often make removal of material and reconstruction of support necessary. In native-valve disease, repair may be possible if it permits complete debridement and reliable function. In right-sided forms, the balance focuses mainly on persistent infection, recurrent pulmonary emboli and poorly tolerated regurgitation, without automatically transferring mitral thresholds and indications. Anatomy and consequences should remain explicit in the decision.

Device extraction is a source-control strategy distinct from valve replacement. It includes planning for complete removal, temporary support if needed and reassessment of the indication for reimplantation. Suppressive antibiotic therapy alone can be a salvage choice when eradication is not achievable, but it is not equivalent to definitive cure. Even in inoperable prosthetic-valve disease, the goals and limitations of suppression should be clear. Defining inoperability requires comparison between procedural risk and the expected course without anatomical source control.

Neurological risk modifies timing in forms with systemic embolization. An ischemic stroke without hemorrhage does not always require delaying a strong cardiac indication; hemorrhage or an infectious aneurysm may require a different sequence. Anticoagulation and antiplatelet therapy are not initiated to prevent vegetation detachment. In recipients of mechanical prostheses, the pre-existing indication is managed separately, balancing prosthetic thrombosis, bleeding and surgical requirements. Prosthetic-valve disease therefore adds a therapeutic problem that native-valve disease may not have.

Prognosis is not a fixed property of the category. Uncomplicated right-sided disease may have favorable prospects for medical cure, but right-sided disease with recurrent emboli, respiratory failure and persistent bacteremia is severe. Similarly, a prosthesis increases complexity without making every treatment futile. Organism, shock, organ injury, extent and feasibility of source control are more precise determinants. Follow-up should retain these data and monitor residual function, recurrences and portals of entry, including healthcare prevention and treatment of substance-use disorder when relevant.

Complications

Mechanical complications vary with the substrate: perforation and chordal rupture in native valves, dehiscence and leaks in prostheses, and annular destruction and fistulas in both. Clinical response to antibiotics may precede or diverge from the evolution of valvular damage. An afebrile patient may therefore require urgent surgery if regurgitation becomes poorly tolerated. This dissociation makes it necessary to follow microbiological control and cardiac function separately.

Embolic dissemination follows blood flow and can create foci capable of sustaining infection. In left-sided forms, the brain and viscera often determine prognosis and timing; in right-sided forms, the lungs and pleura may dominate severity. Fungal infection or a highly mobile vegetation can maintain substantial risk even without the most striking valvular picture. The event should be characterized by site and mechanism because ischemia, abscess and hemorrhage do not require the same management.

Persistence and recurrence are favored by colonized material, abscesses, inadequate antimicrobial exposure and uncontrolled sources. Re-isolation of the same organism requires microbiological comparison and a search for the reservoir; infection with a different organism points more strongly toward a new exposure. In a patient with both a prosthesis and a device, apparent recovery of the valve does not exclude another component as the persisting focus. Anatomical classification should therefore be reassessed when the clinical trajectory does not match the expected response.

Renal failure, toxicity, bleeding and deconditioning occur across all categories and may limit treatment. In frail patients, loss of independence may be as relevant as the microbiological result; in patients with substance-use disorder, lack of continuity after discharge favors new exposures. A final plan should therefore address what remains of the heart disease, what caused the bacteremia and what may hinder adherence. Cure of the episode does not automatically eliminate the predisposition to a subsequent infection.